On Fri, 24 Aug 2007 20:53:03 +0400 Oleg Nesterov <[EMAIL PROTECTED]> wrote:
> (the explicit ack/nack from maintainers is wanted) It's not completely clear who "maintainers" refers to when it comes to this code. > Currently move_task_off_dead_cpu() is called under write_lock_irq(tasklist). > This means it can't use task_lock() which is needed to improve migrating to > take task's ->cpuset into account. > > Change the code to call move_task_off_dead_cpu() with irqs enabled, and change > migrate_live_tasks() to use read_lock(tasklist). > > This all is a preparation for the futher changes proposed by Cliff Wickman, > see > http://marc.info/?t=117327786100003 > > Signed-off-by: Oleg Nesterov <[EMAIL PROTECTED]> > > --- t/kernel/sched.c~1_READ_LOCK 2007-08-12 14:15:37.000000000 +0400 > +++ t/kernel/sched.c 2007-08-24 20:31:03.000000000 +0400 > @@ -5048,7 +5048,7 @@ static void move_task_off_dead_cpu(int d > unsigned long flags; > cpumask_t mask; > struct rq *rq; > - int dest_cpu; > + int dest_cpu, done; > > restart: > /* On same node? */ > @@ -5077,7 +5077,11 @@ restart: > "longer affine to cpu%d\n", > p->pid, p->comm, dead_cpu); > } > - if (!__migrate_task(p, dead_cpu, dest_cpu)) > + > + local_irq_disable(); > + done = __migrate_task(p, dead_cpu, dest_cpu); > + local_irq_enable(); > + if (!done) > goto restart; > } > > @@ -5106,7 +5110,7 @@ static void migrate_live_tasks(int src_c > { > struct task_struct *p, *t; > > - write_lock_irq(&tasklist_lock); > + read_lock(&tasklist_lock); > > do_each_thread(t, p) { > if (p == current) > @@ -5116,7 +5120,7 @@ static void migrate_live_tasks(int src_c > move_task_off_dead_cpu(src_cpu, p); > } while_each_thread(t, p); > > - write_unlock_irq(&tasklist_lock); > + read_unlock(&tasklist_lock); > } > > /* > @@ -5180,11 +5184,10 @@ static void migrate_dead(unsigned int de > * Drop lock around migration; if someone else moves it, > * that's OK. No task can be added to this CPU, so iteration is > * fine. > - * NOTE: interrupts should be left disabled --dev@ > */ > - spin_unlock(&rq->lock); > + spin_unlock_irq(&rq->lock); > move_task_off_dead_cpu(dead_cpu, p); > - spin_lock(&rq->lock); > + spin_lock_irq(&rq->lock); > > put_task_struct(p); > } fyi, with a little rework I queued these changes behind Akinobu Mita's cpu-hotplug changes: cpu-hotplug-slab-cleanup-cpuup_callback.patch cpu-hotplug-slab-fix-memory-leak-in-cpu-hotplug-error-path.patch cpu-hotplug-cpu-deliver-cpu_up_canceled-only-to-notify_oked-callbacks-with-cpu_up_prepare.patch cpu-hotplug-topology-remove-topology_dev_map.patch cpu-hotplug-thermal_throttle-fix-cpu-hotplug-error-handling.patch cpu-hotplug-msr-fix-cpu-hotplug-error-handling.patch cpu-hotplug-cpuid-fix-cpu-hotplug-error-handling.patch cpu-hotplug-mce-fix-cpu-hotplug-error-handling.patch cpu-hotplug-intel_cacheinfo-fix-cpu-hotplug-error-handling.patch cpu-hotplug-intel_cacheinfo-fix-cpu-hotplug-error-handling-fix-a-section-mismatc Lots of other patches keep on coming in and trashing his changes and it's getting a bit bothersome. - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
